AUTOSOMAL HALF-TETRAD ANALYSIS IN DROSOPHILA MELANOGASTERI

M. BALDWIN AND A. CHOVNICK

Department of Genetics. The University of Connecticut, Storrs, Connecticut 06268

Received September 9, 1966

RECOMBINATION studies using tetrad or half-tetrad analyses are particu- larly valuable in investigating the reciprocality of crossover events. Until

recent time, such investigations in Drosophila have been limited to half-tetrad analyses with reversed metacentric compound-X (“attached-X”) chromosomes (see LINDSLEY and SANDLER 1963 €or terminology). Thus, it is possible to con- struct compound-X chromosomes, heterozygous for a series of markers, and to study recombination by examination of the half-tetrad products of meiosis re- covered in the daughters of females carrying such chromosomes. Classic notions concerning crossing over are based upon investigations of this sort (ANDERSON 1925; EMERSON and BEADLE 1935; BEADLE and EMERSON 1935; WELSHONS 1955). In recent years, recombination studies in a number of organisms have elaborated instances of nonreciprocality (see reviews by WESTERGAARD 1964; WHITEHOUSE and HASTINGS 1965). These deviations from the dogma of reciprocality in crossing over share one common feature, namely, that they involve studies of recom- bination between chromosomes labeled with very tightly linked genetic markers, i.e., separable sites within a cistron. Such cases, termed “conversions,” may reflect important features of the process of crossing over, and consequently are the sub- ject of numerous current investigations.

Half-tetrad analysis using compound-X chromosomes suffers from one major defect which makes systematic analysis of conversion in Drosophila melanogaster difficult. and perhaps impossible. Since conversions are expected to arise with low frequency in experiments involving very tight linkage, large scale sampling of meiotic products is essential. Following the synthesis in a single female of a com- pound-X chromosome heterozygous for a series of markers with a specific chro- mosomal distribution, meiosis with recombination will lead to the production of a genotypically heterogeneous population of compound-X females identical in phenotype. Subsequent experiments lwith such females requires single female matings and complete scoring of offspring phenotype in order to determine the chromosomal distribution of markers for each female parent. Such laborious practice seriously limits the usefulness of the compound-X system.

Recently, procedures have been developed for the construction of reversed metacentric compound chromosomes for all autosome arms in Drosophila melano- gaster ( RASMUSSEN 1960; LEWIS, personal communication). These stocks extend

This inrestigation was supported by a research grant, GM 09886, from the Public Iiealth Service

Genetics 5 5 : 277-293 February 1967.

278 M. BALDWIN A N D A. CHOVNICK

half-tetrad analysis to all of the chromosomes of this species, and do not suffer the disadvantage described above for the compound-X system. The major feature of the compound-autosome stocks is illustrated in Figure 1 with compound-third chromosome arms. The figure shows parent females with compound-31 arms and compound-3R arms which are symbolized as C(3L); C(3R). Similarly, the male parent has C (3L’) ; C (3R9. The major classes of male and female prog- eny resulting from union of gametes produced by segregational meioses in both parents will be of two types: C (3L) ; C (3R’) and C (3L’) ; C (3R). The question of nonsegregation is discussed in a later section. If a specific C (3R9 chromosome is constructed carrying a series of heterozygous markers of known chromosomal distribution, then maintenance of the known chromosomal distribution of mark- ers may be accomplished in mass cultures by passing this chromosome from gen- eration to generation only through males. Large numbers of females may be produced, all carrying C(3R’), and these may serve as experimental material for half-tetrad analysis on a mass mating scale. Moreover, if one is able to impose upon the compound-autosome system, a system which selects for survival only progeny carrying interesting half-tetrads, then a useful tool for the systematic investigation of conversion is at hand. Just such an investigation is in progress in this laboratory, and involves half-tetrad analysis of compound-3R chromo- somes. This study assumes that classic rules governing recombination, based upon half-tetrad analysis in single compound-X females, do apply to compound- 3R half-tetrad analysis in mass matings. Consequently, an important preliminary to such an investigation would examine this assumption.

MATERIALS A N D METHODS

The markers used in the half-tetrad analysis and their standard map positions (BRIDGES and BREHME 1944) are curled (cu: 3-50.0), karmoisin (kar: 3-52.0), ebony (e: 3-70.7), and red (red:

FIGURE 1 .-Transmission in Compound-3 stock.

AUTOSOMAL HALF-TETRAD ANALYSIS 279

3-55.5) (OSTER 1954). Inversion (3L)Payne is an inversion in the left arm of chromosome 3 (BRIDGES and BREHME lW), and Cy, SMI is a multiple break rearrangement involving most of chromosome 2 (LEWIS and MISLOVE 1953). Descriptions of other mutants used in this study may be found in BRIDGES and BREHME 194.1.. These mutants are: radius incompletus (ri), peach allele of pink ( p p ) , rosy ( r y ) , glass ( g P ) . The symbolism used to describe compound-autosomes conforms to that adopted for compound chromosomes in general (LINDSLEY, personal communi- cation). Thus, C (3L) RM-P2, ri refers to a specific reversed metacentric compound-31 chromo- some, designated Pasadena 2, and homozygous for the recessive mutant, radius incompletus.

The compound-3R chromosome used in this investigation was constructed by subjecting virgin females of the genotype cu kur red e/++++, to WOr of X-irradiation, and then mating them to males which were C(3L)RM-Peg-i; C(3R)RM-SCI, kar ry . Surviving male progeny of the phenotype, ri kw+, ry+ carried newly synthesized C(3R) chromosomes. A separate stock was established from each male, but the present study was carried out with the derivatives of one such male.

All matings were carried out in half-pint milk bottles containing the standard cornmeal, molasses, yeast, and agar medium seeded with live yeast. Thirty virgin females, 12 to 24 hours old, were mated to 4.0 males of the tester stock, C(3L)RM-P2,ri; C(3R)RM-SBI, ppg l f , and per- mitted to lay eggs for three days. Parents were then transferred to fresh food bnttles, and subse- quent transfers were carried out every two days to provide a total of five broods. Cultures were incubated at 25 * 1°C.

The use of males from the tester stock, C(3L)RM-P&,ri; C(3R)RM-SBl,ppgZs, made it possible to distinguish and discard thosse flies which received a compound-3R from their fathers. All other flies were then scored. All wild-type flies and certain classes of doubles with a recessive phenotype were scored and progeny tested. All other flies with a recessive phenotype were scored and then discarded. Each female to be progeny tested was mated with five males of the tester stock, and her offspring were scored. Each male to be progeny tested was mated to three tester stock virgin females. Two or three of his daughters, carrying the appropriate C(3R) chromosome, were mated in each of two vials with approximately five tester stock males. Because it was possible to test from four to six daughters of each progeny test male, the data gathered for males is generally more complete and thus more accurate than that for females.

Half-tetrad analysis of recombination in the compound-3R chromosome, cu kur red e/++++, was carried out in the presence and absence of heterozygous, heterologous chromosome rearrange- ments. Experiments were conducted with four different female genotypes described below. The abbreviated notation in brackets will he used throughout this report.

1. +/+; C(3L)RM-P2, ri. [+;$I No rearrangements. 2. +/+; C(3L) RM-P5, + . Zn(SL)Puyne, ue h th . +++. [+;Zn(3L)Paym] Inversion

(3L) Payne present.

3. Cy, SMI/+; C(3L)RM-P2, ri [Cy;+] Cy, SMI present.

4. CY, SMI/+; C(3L)RM-P5, +. Zn(3L)Paynq ue h t h . +++. [Cy,Zn(3L)Pnym] Both rearrangements present.

In the sections that follow the reversed metacentric compound autosomes will be referred to as attached-autosomes.

RESULTS A N D DISCUSSION

The general features of recombination elaborated from classical studies with single at tachec-X females may be summarized as follows: I. Crossing over occurs at the tetrad stage, and each exchange involves only two of the four chromatids. 2. Single exchanges occur at random between nonsister chromatids. 3. The fre- quency of homozygosis increases as a function of distance from the centromere. 4. There is no chromatid interference.

280 M. BALDWIN A N D A. CHOVNICK

a b c d (3) +++d

(4) - a b c + a b c +

RECIPROCAL

FIGURE 2.--Single exchanges in C(3R).

NON-RECIPROCAL

Following this model, we may consider t--e recombination resuLLs of mass matings of heterozygous, C (3R) females.

Single exchanges: Exchanges which occur in tetrads of attached-autosomes may be categorized as either reciprocal or nonreciprocal (Figure 2). A recipro- cal exchange is one which occiirs between two nonsister chromatids which are destined to be attached to the same centromere during the second meiotic di- vision, while nonreciprocal exchanges involve nonsister chromatids which will be attached to different centromeres during the second meiotic division.

Both exchange types result in specific half-tetrad products, as illustrated in Figure 2. A reciprocal exchange yields two heterozygous attached-autosomes, one being indistinguishable from a nonexchange attached-autosome (Type-1 ) , and one (which is clearly an exchange attached-autosome (Type-2). In Type-2 there has been an exchange of markers between the two arms to the right of the crossover. On the other hand, a non-reciprocal exchange yields two distinguish- able exchange attached-autosome classes, both of which are homozygous for the markers to the right of the crossover-in one case it is the recessive allele which is homozygous (Type-3) and in the other case the wild-type allele (Type-4).

If crossing over occurs in the tetrad stage, and exchanges occur at random

TABLE 1

Phenotypic frequencies of offspring of C(3R), cu kar red e/++++ mothers with and without heterologous rearrangements

Offspring phenotype

cu kar red e kar red e red e e cu kar red cu kar

kar red red + cu

Heterologous rearrangements in mothers [+ ;+I [+;ln(3L)Puynel [Cy;Zn(3L)Paynel [Cy;+]

50 25 51

266 23

3 1 8 4

2477

49 32 51

244 15 3 0 7 2

2551

51 42 65

290 38 4 0

17 11

2358

32 23 28

131 23 0 1 4

10 1113

Total 2908 2954 2876 1365

AUTOSOMAL HALF-TETRAD ANALYSIS 28 1

between nonsister chromatids, progeny tests of the wild-type offspring should reveal equal numbers of Type-2 and Type-4 flies. It is assumed that failure of a progeny test of wild-type offspring due to death without reproduction occurs at random. This assumption is implied in all ratios examined. Table 1 lists the phenotypic cIasses scored in each experiment. Table 2 presents the results of successful progeny tests carried out on male and female wild-type F,. The num- ber of Type-2 and Type-4 progeny for each region for all four experiments is taken from Table 2 and presented for comparison in Table 4. In all cases, the expected 1 : 1 ratio is found, supporting the conclusion that recombination occurs in the tetrad stage, and takes place at random between nonsister chromatids in the presence as well as the absence oE heterologous rearrangements.

Double exchanges: Figure 3, which has been adapted from a diagram by WELSHONS (1 955), illustrates the possible types of double-exchange products which occur when the first exchange is reciprocal and when it is nonreciprocal. The first exchange is taken as the exchange closest to the centromere. Once the first exchange has been determined, the second exchange can occur in such a way as to form a 2-strand, 3-strand7 or Cstrand double. In the absence of chro- matid interference, 2-, 3-, and 4-strand doubles should occur in a ratio of 1:2: 1.

2-strand

3 - st ram

3-stranc

4-stranc

FIRST EXCHANGE

REG IPROC A L

o b c d (3) - t i t d

(4) (I+++ t b c +

a b c + (7) +++d

FIRST EXCHANGE

NON- RECIPROCAL

a t + d (9) -

a b c d

';= t b c +

a + + d (l3' + + + +

a b c t

FIGURE 3.-Double exchanges in C(3R).

282 M. BALDWIN A N D A. CHOVNICK

TABLE 2

Progeriy test classification of wild-type offspring

cu kar red e

+ + + + + + + + + + + + c u + + + + + + + c.+++ + kar red e

cukar+ + + + + + cukar+ + + + red e cu kar red + + + + + cu kar red + + + + e

+ + + e + + + + + +red e

+ + + + c ~ + + e + + + + c u + + +

+ + + e c u + + e + kar red + c u + + + + kar red + cu kar + e

+ + + + cu kar + + + + + e cukar+ e

1065

21

U)

25

28

35

252

256

15

3

3

2

3

3

4

4

2

3

0

1275

17

15

16

19

21

1 93

189

12

2

3

2

1

0

4

5

0

1

1

964

18

15

13

25

20

221

196

20

1

4

4

1

2

1

4

2

3

0

43 7

10

12

18

28

28

124

120

17

1

5

6

3

3

1

2

3

2

0

820

+ + + + Total 1744 1776 1514

AUTOSOMAL HALF-TETRAD ANALYSIS 283

TABLE 3

Progeny test classification of wild-type male @spring

cu kar red e

+ + + + + + + + + + + + c u + + +

+ + + + c u + + + + kar red e cukar+ + + + + + cu kar + + + + red e cu kar red + + + + + + + + e

+ + + e + + + + + +red e

cu kar red +

+ + + + c u + + e + + + + c u + + + + + + e c u + + e + kar red $-

c u + + + + kar red + cu kar + e

+ + + + cukar+ + + + + e cu kar+ e

+ + red+ cu kar + + + + r e d + + + red + + + + + Total

717

21

15

20

18

28

137

149

15

3

1

0

1

2

4

0

1

0

0

1132

837

17

11

9

8

14

110

98

12

2

1

1

1

0

2

2

0

1

1

588 283

18 10

8 5

6 13

14 21

12 19

1 03 71

95 86

U) 17

1 1

2 3

4 5

1 1

2 1

1 1

4 2

2 2

I 0

0 0

882 541 1127

284 M. BALDWIN AND A. CHOVNICK

TABLE 4

Regional frequencies of Type 2 and Type 4 single-exchange progeny among both male and female offspring (Table 2 )

Heterologous rearrangements Centromere-cu cu-kar kar-red red-e Total

[+;+I Type-2 Type-4 Total

[+ ;In(3L)Payml Type-2 Type-4 Total

Type-4 Total

CCr;+l Type-2 Type-4

Total

CCy;ln(3L)Payml Type-2

25 21 20

4.5 16

17 15 31 13

18 15 28 18

10 12

30

35 28 63 21 19 40 20 25 45 28 28

56

25 6 252 508 189 193 382 196 221 41 7 120 124

244

316 300 616 226 227 4.53 229 261 490 166 164

330

It should be understood that Figure 3 illustrates only one combination of double exchanges between the four gene markers. In the following discussion of double exchanges, all combinations of double exchanges between the gene markers are included in the calculations.

Considering doubles in which the first exchange is reciprocal, there are only two groups of wild-type progeny which are unique in classification. These are a++d/+bc+ and a+++/+bc+ (Types 1 and 4 in Figure 3). The remaining products show a recessive phenotype or would be scored as single exchanges or non-exchanges. If the frequency of 2-strand doubles is equal to one-half the frequency of 3-strand doubles, then Type-1 and Type-4 should occur in a ratio of 1: 1. The first two columns of Table 5 contain the pertinent data, indicating good agreement with the expected 1 : 1 ratio in all experiments.

Now consider those doubles in which the first exchange is nonreciprocal. Types (9) and (13) are identical genotypes, a++d/+f++, as are Types (1 1 ) and (15), a+++/+f+d. Types (9) and (13) will subsequently be referred to as Class A doubles and Types (11) and (15) as Class B doubles. If the frequency

TABLE 5

Summary of doubles found by progeny testing wild-type offspring (Table 2 )

First exchange reciprocal First exchange nonreciprocal

Heterologous rearrangements Type-l TYW-4 Total Class-A Class-B Total ~

5 6 11 7 6 13 14

[Cy; In(3L)PayneI 3 5 8 5 8 13 CCr;+l 6 5 11 6 8 14

Total 15 17 32 25 29 54

~ [+;+I ~

[+; W J L ) P a y m l 1 1 2 7 7

AUTOSOMAL HALF-TETRAD ANALYSIS 285

of 2-strand doubles is equal to the frequency of 4-strand doubles, then Class A and Class B doubles should occur in a ratio of 1: 1. The data (Table 5 , Columns 4 and 5 ) show that 7 Class A and 6 Class B doubles were found in progeny tests of the nonrearrangemerit females. The presence of the heterologous rearrange- ments did not alter this ratio. One further ratio should be considered. Doubles in which the Grst exchange is reciprocal (Type-I +-Type-4) and doubles in which the first exchange is nonreciprocal (Class-A + Class-B) should occur in a ratio of 1:2. Examination of these data in Table 5 indicates that there is no significant difference from the expected ratio in any of the experiments, either singly or pooled.

Consider next the visibly mutant F, (Table 1) which arose from doubles in which the first exchange is nonreciprocal. Inspection of Figure 3 will show that Types (10) and (12) are identical genotypes (abcd/+bcS) as are Types (14) and (1 6) (abc+/+bcd ) . Moreover, all of these show the recessive phenotype bc. Types ( IO) and (12) will be called Class C doubles and Types (14) and (16) are called Class D doubles. Class C and Class D doubles should be found with equal frequency, if the frequency of 2-strand doubles is equal to the frequency of 4-strand doubles. Expand this argument to consider doubles in which the first exchange is nonreciprocal for all combinations of double exchanges in the regions between curled (cu) and ebony ( e ) . Three mutant phenotypic classes are ex- pected (kar, kar red, and red) . Of these, kar offspring were not found (absence of doubles between cu and red) . The remaining two phenotypic classes, kar red and red flies were progeny tested and classified as either Class-C or Class-D doubles. The results of the successful progeny tests of these mutant classes of doubles, summarized in Table 6, indicate that the two classes occur in a 1 : 1 ratio.

Pooling the wild-type (Table 5) and visibly mutant (Table 6) flies which arose from doubles in which the first exchange is nonreciprocal, permits still an- other ratio to be examined. Class A + C and Class B + D doubles should occur with equal frequency, if the frequency of 2-strand doubles is equal to the frequency of 4-strand doubles. The pooled data summarized in Table 7 indicates that a 1 : 1 ratio of these pooled classes was found among the progeny of rearrangement and nonrearrangement mothers.

TABLE 6

Results of progeny testing dmbles with a recessiue phenotype

Class C Class D cu kar red 4- cu kar red 4-

1 letel 11logous rearrangements + kar red + + -k red i- + kar red e i- 4- red e _- _____ cu kar red e cu kar red e

[+;+I 2 2 3 2 I+ Jn(3L)PayneI 3 1 4 1 [Cy; In ( 3 L)Payne] 8 6 7 5 [CY;+] 1 5 3 4

Total 14 14 17 12

286 M. BALDWIN AND A. CHOVNICK

TABLE 7

Pooled data of wild-type and mutant doubles with a nonreciprocal first exchange

Heterologous rearrangements Classes A+C Classes B f D

[+;+I 1 1 1 1 [+;M3L)Pay=l 11 12 [Cy;Zn(3L)Payne] 19 U)

[CY;+] 12 15

Total 53 58

In summary, the analysis of double exchange progeny presented in Tables 5, 6, and 7 shows no indication of chromatid interference,

Calculation of crossouer frequencies: The information in Tables 1 and 2 may be used to estimate crossover frequencies, and these calculations are presented in Table 8. For each region between genetic markers the frequencies were esti- mated in the following manner. The number of chromatids with an exchange in the region of interest is determined from both mutant (Table 1) and wild- type offspring (Table 2). The number of exchanges obtained from the wild-type offspring was adjusted for those progeny tests that failed by multiplying the observed exchanges by the total of wild-type progeny (Table 1) and dividing by the number of successful progeny tests. Thus, for the cu-kar interval in non- rearrangement mothers, the number of exchange chromatids scored in wild-type progeny is 87 (Table 2). Adjusting for the unsuccessful progeny tests leads to an estimate of (87) (2477/1744). To this is added 34 exchange chromatids observed in mutant progeny (Table 1) to yield a total of 157.5 exchange chroma- tids observed in the interval cu-kar in a total of 2908 progeny. Since each offspring represents two chromatids assayed, the frequency of recombination in the inter- val is obtained by 157.5J[ (2908) (2)]. The crossover frequency for the region from the centromere to curled represents a special case, and is obtained in the following manner. The number of crossover progeny among the mutants may be read directly from Table 1. However, due to a bias in the data from wild-type female progeny (discussed in a later section), estimation of the number of wild- type off spring representing crossovers in this region is obtained by considering

TABLE 8

Percent crossing ouer in C(3R) chromosomes

Heterologous rearrangements centromere-cu cu-kar kar-red red-e

[ + ; + I 5.6 2.7 3.9 24.9

[Cy;Zn(3L) Payne] 6.9 2.5 3.6 24.0 [CY;+] 8.3 4.6 6.3 26.6

Standard 2.5* 2.0 3.5 15.2

[+ ;M3L)Paynel 4.7 2.0 2.8 19.1

‘Estimated at one-half the distance between adjacent markers, ri-pp.

AUTOSOMAL HALF-TETRAD ANALYSIS 28 7

the wild-type male progeny tests (Table 3) as the only successful tests. The number of crossovers observed in the male tests is then adjusted for the total of wild-type offspring tested. The sum of mutant crossovers (Table 1) and the adjusted number of wild-type crossovers (Table 3 ) is divided by the total progeny to yield the crossover frequency. Thus for nonrearrangement mothers, the cross- over frequency for the region from centromet-e-cu is estimated as [77 +(39) (2477/1132)]/2908 or 5.6%.

A graphic summary of the recombination data of Table 8 is presented in Figure 4. The percent increase or decrease in recombination (relative to the standard map distances) is plotted for all regions for each of the four genotypes studied. Several points of interest are to be noted in the recombination data. (1) Recombination in the C(3R) chromosome, in the absence of heterologous re- arrangements, is greater in all regions studied than the standard map distances obtained from free-third studies. (2) The presence of the second-chromosome multiple break rearrangement, Cy, S M I , leads to a considerable increase in recombination in the C (3R) chromosome in all regions studied. (3) In contrast, :he presence Zn(3L)Payne in C (3L) is associated with a decrease in recombina- tion in C (3R) in all regions. This decrease may be seen in two sets of compari- sons. (A) Using the recombination data without rearrangements as a standard

I l l I I I I 0 CU kar red IO 20 e

(2.5) (4.5) (8.0) (23.2)

STANDARD THIRD CHROMOSOME

FIGURE 4.-Pattern of recombination in C (3R) with and without heterologous rearrange- ments, compared to standard map distances.

288 M. BALDWIN A N D A. CHOVNICK

TABLE 9

Summary of crossouer and noncrossover chromatids

Noncrossover Single crossover Double crossover

Parents Number Percent Number Percent Number Percent Total

[ + ; + I 2497 71.59 968 27.75 23 0.67 3488 [+ ;I43LPnynel 2833 79.76 708 19.93 11 0.31 3552 [Cy;Zn(3L) Payne] 2251 74.34 761 25.13 16 0.53 3028 [Cr;+l 1082 65.97 535 32.62 23 1.41 16W

compared to the data obtained from mothers carrying Zn(3L)Payne. (B) Using the recombination data in the presence of Cy, SMI as a standard compared to the data obtained with both Cy, SMI and Zn(3L)Payne.

In order to determine how significantly the three rearrangement experiments differed from the nonrearrangement experiment, the data in Table 2 were re- grouped as noncrossover, single crossover, or double crossover chromatids. Each chromatid was considered individually and independently of the others. A sum- mary of these data will be found in Table 9. Inversion (3L)Payne shows a decrease in the classes of single and double crossovers, while Cy, SMI increases these two classes. A homogeneity test was performed between each rearrange- ment experiment and the normal attached-3R, and the results are given in Table I O . The decrease found with In(3L)Payne and the increase !with Cy, SMI are both significant while the distribution of crossover and noncrossover chromatids in the Cy, SMI; In(3L)Payne experiment does not differ from the nonrear- rangement attached3R. Since the female data may introduce a bias into the calculation of homozygosis of wild-type alleles (see discussion below), the data

TABLE 10

Summary of homogeneity tests performed on crossouer and noncrossouer chromatids

Single Double Parents Noncrossover crossover crossover Total P

[ + ; + I 2497 968 E3 3488 [ +;Zn(3L)Payne] 2833 708 11 3552

Total 5330 1676 34 7040 X2 15.52 45.19 4.51 65.22 < 0.01

[ + ; + I 2497 968 23 3488

Total 4748 1729 39 6516

[ICY; Zn(3 L ) Paym] 2251 761 16 3028

X2 1.68 4.20 0.ffi 6.34 0.05

[+;+I 2497 968 23 3488 [CY ;+I 1082 535 23 1 6 M

Total 3579 1503 4 6 5128 X2 5.03 9.02 6.89 20.94 < 0.01

AUTOSOMAL HALF-TETRAD ANALYSIS

TABLE 11

Comparison of the regional distribution of exchange classes

289

Parents centromere-cu cu-kar kar-red red-e Total P -

[+:+I 116 90 137 848 1191 [+ ;Zn(3L)Payne] 99 76 109 679 963

Total 21 5 166 246 1527 21 54 X2 0.16 0.08 0.02 0.04 0.30 0.95

[ + : + I 116 90 137 848 1191 [Cy;Zn(3L)Payne] 132 98 136 814 1180

Total 248 188 273 1662 2371 X2 1.31 0.41 0.00 0.44 2.16 0.50

[+;+I 116 90 137 848 1191 [CY;+] 84 75 103 454 716

Total 200 165 240 1302 1907 X 2 1.69 4.37 2.96 3.97 12.99 0.005

from males alone were classified into noncrossover and crossover chromosomes and tested for homogeneity with the combined data. No difference was found in any of the four experiments.

Another question which may be raised concerns the distribution of exchanges along the marked portion of the attached-3R. Table 11 summarizes homogeneity test comparisons of the crossover class distributions of mutant (Table 1) and wild-type (Table 2) progeny. Taking the progeny of nonrearrangement mothers as a standard, only the Cy, SMl experiment yielded a significant change in the distribution of crossovers.

Homozygosis of recessiue and wild-type alleles: A major expectation of tetrad and half-tetrad analysis is that the percent homozygosis of each marker should increase with distance from the centromere. That this does indeed occur with attached-thirds can be seen in Table 12. The percent recessive homozygosis was calculated by totaling the number of times a recessive marker appeared, and dividing by the total number of flies scored. Thus, in the absence of rearrange- ments in female parents, there are 77 cases listed in Table 1 in which the cu phenotype appears. Since there were 2908 offspring, the percent recessive homo- zygosis for cu is (77/2908) X 100 or 2.65%.

TABLE 12

Percent homozygosis of recessive and wild-type alleles

Parents cu cu+ kar kar+ red red+ e e+

[+;+I 2.6 1.9 3.7 3.1 5.5 4.7 13.5 16.0 [ + ;Zn(3L) Puyne] 2.3 1.6 3.6 2.5 5.3 3.7 12.7 12.0 [Cy;Zn(3L)Payne] 3.2 2.1 5.3 3.4 7.8 4.9 15.6 15.4 [CY;+] 4.1 2.8 6.0 5.1 8.8 8.1 15.7 17.8

290 M. BALDWIN AND A. CHOVNICK

Calculation of the percent homozygosis for the wild-type allele is more diffi- cult. Considering the nonrearrangement females in Table 2 there are 39 cases of cu+ in 1744 successful progeny tests. Adjusting for those lwild-type flies which were unsuccessful in progeny tests, the percent homozygosis for CUI- among all flies is estimated by (39/1744) X (2477/2908) x 100 or 1.9%. Table 12 presents the percent homozygosis for both wild-type and recessive alleles.

It is evident from Table 12 that the percent homozygosis for the wild-type alleles is lower than that for the recessive alleles of those markers close to the centromere. The only explanation which can be offered for this discrepancy is that a bias was introduced by using the data gathered from testing wild-type females. Since each female produced relatively few progeny, determination of the genotype of a phenotypically wild F, female was often impossible. This difficulty is greater for markers closer to the centromere. In contrast, the data from wild F, males is more reliable. Each such male was mated, and several of his daughters were progeny tested to determine his genotype. Thus, the size of sample of progeny tests of each wild-type male permitted easy classification. In order to determine whether the data from females introduced a bias into the results, the male data alone (Table 3) were used to calculate the percent homo- zygosis of !wild-type alleles. These results, presented in Table 13, show close agreement for homozygosis of wild-type and recessive alleles, and support the contention that the female data did bias the results.

Meiotic segregation of C(3L) and C(3R) chromosomes: In the introduction, discussion of transmission of attached-autosomes was restricted to a considera- tion of the classes of offspring produced by gamete unions resulting from meiotic segregation of C (3L) and C (3R) chromosomes (Figure 1 ) . In two of the present experiments the C (3L) and C (3R) chromosomes were differentially marked in the parents, thus providing an opportunity to observe segregation. These data are presented in Table 14.

In both cases the male parent was the tester stock in which C(3L) is marked by the mutant phenotype, radius incompletus ( r i ) , and C(3R) is homozygous for p p gP. The female parent in both experiments carried the C (3L) heterozygous for Zn(3L)Payne and was ri+, while the C(3R) chromosome was the subject of the recombination analysis. The two experiments differed in that the females, in one experiment, were heterozygous for the second chromosome rearrange- ment, Cy,SMI. Several features of interest are to be noted in Table 14. In ad-

TABLE 13

Percent homozygosis of wild-type alleles using data from males only

Parents

cu+

Number Percent

kar+ red+ e+

Number Percent Number Percent Number Percent

55 4.1 74 5.6 193 14.5 4.5 3.4 54 4.1 14.8 11.3 53 4.9 71 6.6 146 13.6 41 6.2 64 9.7 10s 16.3

AUTOSOMAL HALF-TETRAD ANALYSIS 29 1

TABLE 14

Segregation of C(3L) and C(3R) chromosomes in crosses of C(3L), + . In(3L)Pape; C(3R), cu kar red e/+++-1- females to males of the tester stock, C(3L),ri; C(3R),pP g13

Segregational offspring Nonsegregational offspring Chromosome 2 - in mothers Brood No. ri+; p p glJ ri; ( p p gP)+ ri+; ( p p gP)+ ri; p p gB Total

-t/+ 1 672 702 12 4 1390 2 526 553 3 4 1086 3 428 496 6 1 931 4 672 688 3 4 1367 5 465 487 4 3 959

Total 2763 2926 28 16 5733

CY, SMI /+ 1 338 447 83 22 890 2 423 555 113 32 1123 3 467 553 64 31 1115 4 356 4.49 52 31 888 5 378 514 45 16 953

Total 1962 2518 357 132 4969

dition to the two classes of progeny reflecting meiotic segregation of the C (3L) and C (3R) chromosomes, both experiments yielded two classes of off spring which could only result from nonsegregational meioses in both parents. Thus, the ri+(ppgZ3)+ class are matroclinous for both C(3L) and C(3R), and the ri, p”, gZ3 class are patroclinous for these chromosomes. Of particular interest is the more than tenfold increase in the frequency of nonsegregational off spring occasioned by the introduction of Cy, S M l into the maternal genotype. Assum- ing a constant frequency of nonsegregational meioses in the male parent com- mon to both experiments, and in view of the fact that the introduction of Cy, S M l to the maternal genotype led to no significant drop in fertility, it is sug- gested that there is a high frequency of nonsegregational meioses in the attached- 3 males. Since nonsegregational progeny were not reported in other extensive experiments (MCCLOSKEY 1966), it is inferred that nonsegregational meioses in females is rare in the absence of chromosome rearrangements.

Examination of the distribution of nonsegregational off spring in both experi- ments reveals two other points of some interest. In both experiments the matro- clinous progeny occur with greater frequency, and in the Cy, S M l experiment the discrepancy is most significant. The second point concerns homogeneity of the data on frequency of nonsegregational offspring in successive broods. In both experiments the patroclinous progeny frequencies are quite homogeneous. In contrast, there is a deviation from homogeneity for the matroclinous progeny which assumes statistical significance in the C y , S M l experiments. Examination of the brood distribution of matroclinous progeny reveals a much greater fre- quency in early broods than in later broods. Explanation of the discrepancy between the two classes of nonsegregational progeny on the basis of decreased viability associated with the C(3L) chromosome marked with ri, and the C(3R)

292 M. BALDWIN AND A. CHOVNICK

chromosome marked with p p gP will not satisfactorily explain the extreme dis- crepancy in the Cy, SMI experiments. Moreover, in other experiments involv- ing other marker systems (HOLM and CHOVNICK, unpublished), the non- segregational progeny exhibit a similar discrepancy in the same direction. The authors suggest that these features of the distribution of nonsegregational progeny may best be understood with the following suppositions: (1) Nullo-3 eggs are uniformly less viable or are formed with a lower frequency than C(3L) ; C(3R) eggs. Either of these suppositions will explain the discrepancy between the two classes of nonsegregational progeny, and is consistent with the uniform brood distribution of patroclinous offspring. ( 2 ) In addition, the sharp decrease in frequency of matroclinous progeny in successive broods may be understood on the additional assumption that nullo-3 sperm do not store well.

The suggestions of lowered viability of nullo-3 eggs, and lowered capacity for storage of nullo-3 sperm, imply gene function in these cell types. These sugges- tions are in conflict, at least for sperm, with conclusions drawn earlier (MULLER and SETTLES 1927; MCCLOSKEY 1966), and further investigations bearing on these questions are in progress.

Question may be raised concerning the relationship, if any, of the nonsegre- gational meioses in females with the change in crossover frequencies seen in C(3R) in the presence and absence of Cy, SMI and Zn(3L)Payne. Unfortu- nately, the present data are inadequate, since segregation data could not be col- lected for those experiments in which Zn(3L)Payne was absent. In those experi- ments, both parents possessed the same C(3L) chromosome marked with ri. Moreover, in those experiments for which segregation data was collected, no effort was made to collect the recombination data such that it might be related to the segregation data. Consequently, it is not possible to compare the frequency of crossover chromatids among matroclinous progeny with those from segrega- tional progeny.

SUMMARY

The principles of recombination established by half-tetrad analysis of single attached-X females also hold for attached-3R chromosomes, and these principles may be elaborated from mass matings of attached-3R females, The following observations demonstrate this similarity in C (3R) (Compound 3R) : (1 ) Re- ciprocal and nonreciprocal single exchanges occm with equal frequency. ( 2 ) In the case of double exchanges there is no evidence of chromatid interference. That is, the second exchange in a double exchange tetrad is independent of the first. (3) Percent homozygosis of both wild-type and mutant alleles increases with the distance from the centromere. (4) The presence of the heterologous rearrangements, Cy, SMI or Inversion(3L)Payne or both, does not alter either the ratio between reciprocal and nonreciprocal single exchanges or the ratio between 2-strand, %strand, and 4-strand doubles. ( 5 ) The presence of Cy, SMI was shown to increase recombination along the marked portion of the attached- 3R. However, crossover values were decreased in the presence of Inversion (3L) Payne, while both together increased it only slightly.-Observations on meiotic

AUTOSOMAL HALF-TETRAD ANALYSIS 293

segregation of the C (3L) and C (3R) chromosomes indicate that: (1) Nonsegre- gational meioses take place regularly in males with a high frequency (not esti- mated in the present study). It is inferred from other studies (MCCLOSKEY 1966) that in the absence of chromosome rearrangements, nonsegregational meioses are rare in females. (2) Observations on the relative frequency of matroclinous and patroclinous progeny suggest either that nullo-3 eggs are produced with a lower frequency than C (3L) ; C (3R) eggs, or that nullo-3 eggs are uniformly less viable than C (3L) ; C (3R) eggs. (3) Observations on the brood distribution of matro- clinous progeny suggest that upon storage, nullo-3 sperm are less likely to lead to successful fertilization.

LITERATURE CITED

ANDERSON, E. G., 1925 Crossing-over in a case of attached-X chromosomes in Drosophila

BEADLE, G. W., and S. EMERSON, 1935 Further studies of crossing-over in attached-X chromo-

BRIDGES, C. B., and K. S. BREHME, 1944 The mutants of Drosophila melanogaster. Carnegie

EMERSON, S., and G. W. BEADLE, 1935 Crossing-over near the spindle fiber in attached-X

LEWIS, E. B.. and R. F. MISLOVE, 1953 Reports of new mutants. Drosophila Inform. Serv. 27: 58.

LINDSLEY, D. L., and L. SANDLFX, 1963 Construction of the compound-X chromosomes in Drosophila melanogaster by means of the Bar-Stone duplication. pp. 39043. Methodology in Basic Genetics. Edited by W. E. BURDETTE. Holden-Day, San Francisco.

The problem of gene activity in the sperm of Drosophila melanogaster. Am. Naturalist 100: 21 L-218.

The non-functioning of the genes in spermatozoa. Z. Ind.

melanogaster. Genetics 10: 4 0 3 4 1 7.

somes of Drosophila melanogaster. Genetics 20 : 192-206.

Inst. Wash. Publ. 552.

chromosomes of Drosophila melanogaster. Z. Ind. Abst. Vererb. 65: 129-14.0.

MCCIBSKEY, J. D., 1966

MULLER, H. J., and F. SETTLES, 1927 Abst. Vererb. 43: 285-312.

OSTER, I. I., 1954 Reports of new mutants. Drosophila Inform. Sew. 28: 77-78. WELSHONS, W. J., 1955

WESTERGAARD, M., 1964.

WHITEHOUSE, H. L. K., and P. J. HASTINGS, 1965

A comparative study of crossing-over in attached-X chromosomes of

Studies on the mechanism of crossing-over. I. Theoretical considera-

The analysis of genetic recombination on

Drosophila melamgaster. Genetics 40 : 91 8-936.

tions. Compt. Rend. Trav. Lab. Carlsberg 34: 359405.

the polaron hybrid DNA model. Genet. Res. 6: 27-92.